I have 2 years of daily data of demand of some product. I can observe daily as well as yearly seasonality. I did run decompose method(frequency = 365) to separate data into trend, seasonal and random components. I am interested in getting random data but it gives almost constant random component.

Then I did use TBATS model with multiple frequency (7,365) and getting level,scope,season1,season2 components separately. Can i get the random component from this ?

Actual problem is I want to find significance of one variable on demand but seasonality dominates that and i am not getting any correlation. So i am trying to find correlation from the random data ...


The "random" component in a decomposition is the remainder from subtracting trend+season from y. So in the TBATS model, the residuals are the equivalent.

  • $\begingroup$ Thank U very much ... plz clarify one thing, I want to find price change effect on the random data ( Residuals(tbats)) .. In the whole data, i am getting only seasonality effect ....Am i going in right way ?? $\endgroup$ – user1590205 May 13 '17 at 10:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.